Request a Report Request a Demo

Consumer Data and Its Applications in Health Insurance Underwriting

Can consumers’ purchasing habits really help insurers to see group health risk more clearly? This study found that adding consumer data to our Curv® predictive model yielded almost no incremental lift.

Curv® disrupted group health insurance over ten years ago by using deidentified prescription and medical claims data to score risk on small to mid-sized groups with little to no claims histories.

So if some data is valuable, is more data always better? Turns out it isn’t.

  • A team of actuaries and data scientists studied the impact of adding consumer data features to Curv.
  • The study looked at approximately 175,000 groups of five to 500 employees—a total of about 14 million individuals.
  • The consumer dataset included several hundred features in categories such as purchase history (e.g., house, car), consumer preferences (e.g., magazine subscriptions, shopping habits), and credit attributes (e.g., income, mortgage size, payment history).

Despite the fact that consumer data improves lift in other categories, the authors found that the consumer data they added to the Curv model provided almost no incremental lift in group health or stop-loss insurance.

The authors concluded that their study results are “stark enough that we believe skepticism is warranted towards the use of consumer data in group health underwriting applications.”

Please enter your email to keep reading. Don’t worry—we promise we won’t start hounding you with calls and emails. We just need to make sure our valuable content is delivered to the right audience.